/**
 * TextAnalyzer.java
 * Boston University CAS CS 112 HW5
 * By Leo Reyzin
 */

import java.util.*;
import javax.swing.*;


public class TextAnalyzer {

	private CountingTree <String> ctree;
	private BSTWithDuplicates<Integer, String> ftree;

	
	/*
	 * Reads words from fileScanner, starting at the first occurrence
	 * of firstWord (inclusive) until the first occurrence of endWord
	 * after that (exclusive).  If firstWord is the empty string,
	 * starts at the beginning of the file.  If
	 * endWord is the empty string, goes until the end
	 * of the file.
	 * 
	 * Builds a binary search 
	 * tree indexed by words that has a count of how many times
	 * each word occurs, and another binary search tree indexed by
	 * count that has all the words that appear a given number of times.
	 * 
	 * Returns a string to be output to the user.
	 */
	public String readText (Scanner fileScanner, String firstWord, String endWord) {
		String ret = "";
		
		// This will make sure we read only words, no spaces, numbers, punctuation, etc.
		// Unfortunately, apostrophes and hyphens are considered delimiters
		// here (have to be, in order for quotes and dashes to work correctly)
		// which means that words with apostrophes and hyphens get split,
		// rather than taken as single words 
		fileScanner.useDelimiter("[^a-zA-Z]+");

		String word = null;

		// Skip all the text until firstWord appears
		if (!firstWord.equals("")) {
			boolean found = false;
			while (fileScanner.hasNext()) {
				word = fileScanner.next();
				if (word.equals(firstWord)) { // found the first word -- don't forget to insert it into the tree below
					found = true;
					break;
				}
			}

			if (!found) {
				ret+="First word not found.\n";
				ctree = null;
				ftree = null;
				return ret;
			}
		}
		
		// first word is empty, so we need to start reading at the beginning
		else {
			if (!fileScanner.hasNext()) {
				ret+="No words found.\n";
				ctree = null;
				ftree = null;
				return ret;
			}
			word = fileScanner.next(); // read the first word -- it has to be treated separately, because the code when firstWord is not empty will read it
		}

		// Insert all of the words in to the tree, converting them to lowercase first
		ctree = new CountingTree<String>();
		boolean found = false;
		boolean endWordExists = !endWord.equals("");

		ctree.insert(word.toLowerCase()); // insert the first word		
		while (fileScanner.hasNext()) {
			word = fileScanner.next();
			if (endWordExists && word.equals(endWord)) {
				ret+="End word found.\n";
				found = true;
				break;
			}
			ctree.insert(word.toLowerCase());
		}
		if (endWordExists && !found) {
			ret+="End word NOT found; end of file reached.\n";
		}
		ret+="\nTotal Words: "+ctree.getTotalEntries()+".";
		ret+="\nDistinct Words: "+ctree.getDistinctEntries()+".";

		ftree = ctree.frequencyTree();
		ret+="\nMost Frequent Word Appears "+ftree.maxKey()+" times.";
		ret+="\nYou may now search for all the words appearing within a range of frequencies.";
		return ret;
	}
	
	/*
	 * Returns a string containing, one per line, all the words that appear
	 * between low and high number of times in the analyzed text.
	 * First line of the strings tells how many such words have been found.
	 * Next to each word is its frequency.
	 * The string is meant to be output to the user.
	 * If readText has not been run or has not found any words in the text,
	 * returns an error message.
	 */

	public String printWords (int low, int high)
	{
		if (ftree == null)
			return "No text has been analyzed.\n";

		String ret = "";
		Collection<String> result =  ftree.rangeSearch(low, high);
		ret+="Found "+result.size()+" words.\n";
		for (String s:result) {
			// looking up the frequency of s in ctree is not the most
			// efficient way to do this -- it would be more efficient
			// to have the collection contain (word,frequency) pairs
			// But that would slow down the rangeSearch, and the user
			// may not always want to print out the results, anyway.  Besides,
			// the extra time for lookup is small compared to user interaction
			// time.
			ret+=String.format("%5d %s\n",ctree.search(s),s);
		}
		return ret;
	}



}